When translation becomes a crutch: rethinking L1 reliance in second language classrooms

Cuando la traducción se convierte en una muleta: replanteando la dependencia de la L1 en las aulas de lenguas extranjeras

          Adonay Bárbara Pérez Luengo[1]

             Manuel de Jesús Velázquez León[2]

Maria de la Caridad Smith Batson[3]

Abstract

Teachers of second languages routinely face a practical tension: translation to learners’ L1 is fast, accurate, and reassuring—but is it good for acquisition? Building on nine years of teaching English and Spanish to Mandarin-L1 university students in China, this paper argues that frequent translation, especially when treated as the goal of listening or reading, can impede second language acquisition (SLA) by short-circuiting core psycholinguistic processes (input processing, form–meaning mapping, noticing, automatization) and by reinforcing lexical routes that keep L2 dependent on L1 mediation. Drawing on SLA theory (comprehensible input, interaction, noticing, input processing), bilingual memory models (the Revised Hierarchical Model), and cognitive load theory, we review evidence on classroom L1 use, code-switching/translanguaging, dictionary and gloss studies, and extensive reading/listening. The review supports a calibrated stance: low-frequency, purpose-bound L1 use can be facilitative (e.g., form-focused contrast, task setup, affective support), while habitual translation as comprehension produces shallow processing, low tolerance for ambiguity, and slower growth of direct L2–concept links. We synthesize these findings into a practical “Comprehension-First, Translation-Lite” pedagogy, including routines for monolingual support scaffolds, graduated glossing, noticing-rich tasks, and assessment designs that reward L2-to-concept processing. The paper concludes with actionable recommendations for instructors and program designers.

Keywords: Second Language Acquisition (SLA), translation, L1 Mediation.

Resumen

Los profesores de segundas lenguas se enfrentan habitualmente a una disyuntiva práctica: traducir a la lengua materna de los estudiantes es rápido, preciso y tranquilizador, pero ¿es beneficioso para la adquisición del idioma? Basándonos en nueve años de experiencia docente con estudiantes universitarios chinos, cuya lengua materna es el mandarín, este artículo sostiene que la traducción frecuente, especialmente cuando se considera el objetivo de la comprensión auditiva o lectora, puede obstaculizar la adquisición de segundas lenguas (ASL) al interferir con procesos psicolingüísticos fundamentales y al reforzar rutas léxicas que mantienen la L2 dependiente de la mediación de la L1. Apoyándonos en la teoría de la ASL, los modelos de memoria bilingüe y la teoría de la carga cognitiva, analizamos la evidencia sobre el uso de la L1 en el aula, la alternancia de códigos/translingüística, los estudios de diccionarios y glosas, y la lectura/audición extensiva. La revisión respalda un enfoque equilibrado: el uso poco frecuente y específico de la L1 puede ser facilitador, mientras que la traducción habitual como comprensión produce un procesamiento superficial, baja tolerancia a la ambigüedad y un desarrollo más lento de las conexiones directas entre la L2 y los conceptos. Sintetizamos estos hallazgos en una pedagogía práctica de «Prioridad a la Comprensión, Traducción Ligera», que incluye rutinas para el apoyo monolingüe, glosado gradual, tareas que fomentan la observación y diseños de evaluación que premian el procesamiento de la L2 a conceptos. El artículo concluye con recomendaciones prácticas para docentes y diseñadores de programas.

Palabras clave: adquisición de una segunda lengua, traducción, mediación de la L1

Introduction

In many university classrooms in China and elsewhere, translation is omnipresent. When students cannot follow a recording, they ask for a Chinese gloss; when reading, they translate sentences line-by-line; when studying vocabulary, they store bilingual pairs; and when assessed, they often equate success with the accuracy of L2-to-L1 paraphrases. Over nine years teaching English and Spanish to Chinese undergraduates, we have repeatedly observed the belief that translation is comprehension—and, by extension, that successful language learning consists largely of rapidly converting L2 forms into their L1 equivalents.

This paper questions that belief. The thesis is deliberately narrow: frequent translation—especially when it becomes the default route to understanding—can hinder L2 acquisition. The claim does not demonize L1 use; rather, it distinguishes targeted, low-frequency uses (e.g., brief contrastive notes, affective support) from habitual, end-state translation that supplants L2-internal processing. Acquisition here refers to durable, automatized ability to map L2 forms to concepts and to comprehend and produce in real time, not merely to translate.

This orientation aligns with comprehension-based theories in SLA, which emphasize meaning-focused input and interaction as central drivers of development when processing conditions are right (Krashen, 2004; Long, 1996; Nakanishi, 2015). At the same time, the persistence of grammar-translation traditions in parts of China helps explain why undergraduates arrive at university with entrenched habits of sentence-by-sentence translation (Liu, 2021).

Development

Theoretical framework: why habitual translation can short-circuit acquisition

Comprehensible input remains a central condition for acquisition, with decades of research on reading- and listening-driven growth demonstrating benefits for vocabulary, fluency, and general proficiency when input is meaningful, abundant, and appropriately leveled (Elley, 1991; Krashen, 2004; Nakanishi, 2015). Interaction and negotiated meaning further enhance intake by focusing attention and providing feedback, while productive demands can push learners to restructure interlanguage (Long, 1996). If learners routinely route comprehension through L1 translation rather than directly building L2 meaning, they risk weakening the very processes that drive acquisition.

Noticing and input processing are particularly germane. Schmidt’s (1990) Noticing Hypothesis proposes that learners must consciously notice certain features for input to become intake. VanPatten (1996) argues that learners prioritize meaning over form and require conditions that help them map forms to meanings during comprehension. Translation that provides ready-made L1 equivalents can reduce the need to notice form cues, thereby bypassing the mapping work that builds L2 representations.

Cognitive Load Theory reminds us that working memory is limited; pedagogies should minimize extraneous load so learners can allocate resources to germane processing (Sweller, 1988). Translation can cut two ways: as a shortcut it may reduce immediate load, but as a dual-route operation—holding L2 forms while constructing L1 equivalents—it can also add extraneous steps that displace deeper L2 semantic processing. Relatedly, research on levels of processing suggests that shallow, surface mapping (word-for-word translation) yields weaker memory than deeper semantic integration (Craik & Lockhart, 1972).

Bilingual lexical architecture offers another lens. The Revised Hierarchical Model posits asymmetric links between L1/L2 lexical and conceptual stores: early L2 processing relies on L2-to-L1 lexical links; with proficiency, direct L2-to-concept mappings strengthen (Kroll & Stewart, 1994). Overreliance on translation may maintain the weaker, indirect route (L2→L1→concept) rather than fostering direct conceptual access in L2. From a skill acquisition perspective, fluent performance emerges as declarative knowledge proceduralizes through practice under appropriate conditions (DeKeyser, 2012). If comprehension consistently depends on translation, learners practice the wrong micro-skill (conversion), not the target one (real-time L2 parsing), slowing automatization.

Literature review

The last two decades have moved from a prescriptive target language only dogma toward judicious L1 use. Turnbull and Dailey (2009) gathered cases where L1 played useful roles (explaining complex grammar, task management, affective support) and advocated an optimal rather than zero-use stance. Macaro (2009) likewise argued for principled code-switching as a communication and learning strategy. Yet even these calibrated positions distinguish between strategic L1 use and pervasive reliance; studies note risks of dependency, reduced exposure to L2, and missed opportunities for negotiation of meaning.

Translanguaging scholarship reframes bilingual practices as dynamic repertoires, challenging rigid separation of languages and documenting pedagogical benefits (García & Li Wei, 2014; Li Wei, 2018). These works provide strong arguments for leveraging the whole linguistic repertoire—but they do not equate translanguaging with constant word-level translation as the goal of comprehension. Treating L1 translation as the endpoint of reading/listening is a narrower practice with different cognitive consequences than translanguaging’s broader, agentive meaning-making.

A robust empirical strand investigates L1 versus L2 glosses and dictionary types. Findings are nuanced: L1 glosses often aid immediate comprehension and sometimes retention of individual word forms; monolingual resources can foster deeper definitional knowledge and production (Laufer & Hadar, 1997; Nation, 2001; Yoshii, 2006). Particularly relevant is Laufer and Girsai’s (2008) demonstration that contrastive analysis plus translation can benefit vocabulary learning in form-focused instruction. The implication is not that translation should be ubiquitous, but that targeted L1-mediated contrasts can sharpen form–meaning mappings when the objective is explicit lexical learning.

Meta-analyses and longitudinal book flood studies consistently tie extensive input to proficiency gains across domains (Elley, 1991; Nakanishi, 2015). These interventions acculturate learners to build meaning within the L2, with minimal recourse to translation, cultivating tolerance for ambiguity and faster parsing. In the Chinese EFL ecosystem, however, the persistence of grammar-translation practices remains documented, especially in exam-oriented settings (Liu, 2021).

Practitioner Evidence: vignettes from Mandarin-L1 Classrooms

Classroom observations align with these theoretical and empirical patterns. In a B1 English listening class, students initially insisted on Chinese summaries after each segment. When, for two weeks, such summaries were replaced with English paraphrase chains plus visual support and comprehension checks, immediate quiz scores dipped, but by week three, notes improved and students reported hearing more English before wanting Chinese. The shift resembled moving from L2-to-L1 mediation to L2-to-concept paraphrasing, allied with noticing gains evident in their annotations. In a Spanish news reading course (A2→B1), two groups alternated supports: bilingual word glosses versus monolingual simplified definitions with example sentences and pictures.

The bilingual group finished faster but needed more dictionary lookups on a delayed recall test; the monolingual group worked more slowly but retained collocations and could paraphrase headlines in Spanish. Brief, explicit contrastive notes (L1) effectively prevented fossilized misuse of false friends (e.g., actual vs. real), but were followed by English-only example generation to build direct L2 connections. On high-anxiety days, short L1 check-ins reduced stress and improved participation; however, when open L1 discourse was allowed during task time, students reverted to negotiating ideas in Chinese and left English for surface reporting.

Mechanisms of hindrance

Several mechanisms can explain how frequent translation hinders acquisition. First, ready L1 equivalents bypass the need to map forms to meanings, weakening the form–meaning links that input processing seeks to strengthen (VanPatten, 1996). Second, according to the Revised Hierarchical Model, persistent L2-to-L1 reliance maintains indirect conceptual access and delays the strengthening of direct L2-to-concept links (Kroll & Stewart, 1994). Third, serial L2-to-L1 conversion imposes extraneous working-memory load (Sweller, 1988), leaving fewer resources for noticing and integration. Fourth, word-for-word translation fosters shallow processing, whereas deeper semantic integration yields more durable memory (Craik & Lockhart, 1972). Fifth, when translation is always available, learners are less likely to engage in negotiation of meaning and pushed output, missing feedback that drives development (Long, 1996). Finally, assessment washback matters: if quizzes reward L2-to-L1 accuracy rather than L2-internal comprehension, students rationally adopt translation as their primary study strategy, entrenching the cycle.

A Comprehension-First, Translation-Lite Pedagogy

A comprehension-first, translation-lite pedagogy repositions L1 as strategic scaffold rather than default endpoint. Input design should invite L2-internal sense-making through compelling extensive reading/listening with levelled texts and clear post-tasks that reward message comprehension rather than bilingual equivalents (Elley, 1991; Krashen, 2004; Nakanishi, 2015). Pre-comprehension supports should be provided primarily in the L2 via visuals, synopsis sentences, and preview questions, keeping pre-teaching vocabulary minimal and semantic.

A graduated support ladder can guide resource choice: Stage 1, L2 paraphrases and pictures (default); Stage 2, bilingual glosses for truly low-frequency, high-yield items with immediate return to L2 use; Stage 3, contrastive notes for stubborn false friends or collocational traps, followed by practice in L2 contexts (Laufer & Girsai, 2008). Noticing-rich tasks such as text enhancement and dictogloss can make form matter without sacrificing meaning (Schmidt, 1990; VanPatten, 1996). Interactional protocols should encourage clarification requests and circumlocution before resorting to L1. Finally, assessment must reward L2-internal processing: comprehension checks via L2 paraphrase, inference, and form–meaning mapping items, with translation items kept rare and purpose-specific.

Counterarguments and responses

Common counterarguments deserve careful consideration. One claim is that translation speeds learning. This is often true for immediate comprehension and selected vocabulary targets, but speed can be purchased at the cost of depth and transfer; delayed tests often show monolingual supports yield better L2 production and collocational accuracy (Laufer & Hadar, 1997; Yoshii, 2006).

Another claim is that translanguaging is empowering and that restricting L1 is inequitable. The position here is not to prohibit translanguaging, but to distinguish it from habitual sentence-level translation as a task goal; the former can democratize participation while the latter can short-circuit acquisition processes (García & Li Wei, 2014; Li Wei, 2018). Finally, beginners are said to need translation. At very low proficiency, micro-doses of translation prevent paralysis, but even beginners can perform L2-internal tasks with pictorial support, gestures, and graded input; a laddered support system addresses precisely these stages.

The recommendations are that programs and classrooms can operationalize this stance through several moves. First, articulate a program-level policy on L1 use: strategic, limited, purpose-bound L1; comprehension outcomes assessed in the L2. Second, guarantee sustained extensive reading and listening with levelled materials and L2-internal post-tasks. Third, invest in teacher development around input-processing tasks, text enhancement, and negotiation moves, paired with clear protocols for when and how to deploy L1 briefly and how to exit the L1 red button. Fourth, reform assessment by reducing translation items, and adding L2 paraphrase, inference, and form–meaning mapping tasks. Fifth, induct students early: teach why translation is not comprehension, visualize the Revised Hierarchical Model, and model paraphrase chains. Finally, replace bilingual wordlists with monolingual learner-friendly definitions plus examples, reserving bilingual glosses for a small subset of items where contrastive precision is essential.

Conclusions

Frequent translation is tempting because it is fast and certain, but acquisition is neither; it is slow and probabilistic, driven by countless episodes of L2-internal meaning-making in which forms are mapped to concepts, noticed in context, and practiced to fluency.

Theories of input processing, noticing, cognitive load, bilingual lexical architecture, and skill acquisition, all predict that habitual translation as the endpoint of comprehension, can impede these processes.

The empirical record supports a calibrated stance: L1 as a scalpel, not a hammer—useful for targeted contrast, quick clarifications, and affective care; counterproductive when it becomes the default.

A Comprehension-First, Translation-Lite model offers practical routines and assessment practices suited to Chinese university contexts and beyond.

References

Craik, F. I. M., & Lockhart, R. S. (1972). Levels of processing: A framework for memory research. Journal of Verbal Learning and Verbal Behavior, 11(6), 671–684.

DeKeyser, R. (2012). Automatization, skill acquisition, and practice in second language acquisition. In C. A. Chapelle (Ed.), The encyclopedia of applied linguistics. Wiley.

Elley, W. B. (1991). Acquiring literacy in a second language: The effect of book-based programs. Language Learning, 41(3), 375–411.

García, O., & Li Wei, L. (2014). Translanguaging: Language, bilingualism and education. Palgrave Macmillan.

Kroll, J. F., & Stewart, E. (1994). Category interference in translation and picture naming: Evidence for asymmetric connections in bilingual memory. Journal of Memory and Language, 33(2), 149–174.

Krashen, S. D. (2004). The power of reading: Insights from the research (2nd ed.). Libraries Unlimited.

Laufer, B., & Girsai, N. (2008). Form-focused instruction in second language vocabulary learning: A case for contrastive analysis and translation. Applied Linguistics, 29(4), 694–716.

Laufer, B., & Hadar, L. (1997). Assessing the effectiveness of monolingual, bilingual, and bilingualised dictionaries in the comprehension and production of new words. The Modern Language Journal, 81(2), 189–196.

Li Wei, L. (2018). Translanguaging as a practical theory of language. Applied Linguistics, 39(1), 9–30.

Liu, P. (2021). Grammar-translation and its use in present-day China. Journal of Arts & Humanities, 10(2), 1–9.

Long, M. H. (1996). The role of the linguistic environment in second language acquisition. In W. Ritchie & T. Bhatia (Eds.), Handbook of second language acquisition (pp. 413–468). Academic Press.

Macaro, E. (2009). Teacher use of code-switching in the L2 classroom: Exploring ‘optimal’ use. In M. Turnbull & J. Dailey-O’Cain (Eds.), First language use in second and foreign language learning (pp. 35–49). Multilingual Matters.

Nakanishi, T. (2015). A meta-analysis of extensive reading research. TESOL Quarterly, 49(1), 6–37.

Nation, I. S. P. (2001). Learning vocabulary in another language. Cambridge University Press.

Schmidt, R. (1990). The role of consciousness in second language learning. Applied Linguistics, 11(2), 129–158.

Sweller, J. (1988). Cognitive load during problem solving: Effects on learning. Cognitive Science, 12(2), 257–285.

Turnbull, M., & Dailey-O’Cain, J. (Eds.). (2009). First language use in second and foreign language learning. Multilingual Matters.

VanPatten, B. (1996). Input processing and grammar instruction in second language acquisition. Ablex.

Yoshii, M. (2006). L1 and L2 glosses: Their effects on incidental vocabulary learning. Language Learning & Technology, 10(3), 85–101.

[1] Licenciado en Educación, especialidad Inglés. Máster en Ciencias de la Educación. Doctora en Ciencias Pedagógicas. Profesora Titular en Universidad Shantou. Conferencista Facultad de Artes Liberales. Guangdong, República Popular China, Correo electrónico: adonay@stu.edu.cn registro ORCID (https://orcid.org/0000-0002-1513-850X).

[2] Licenciado en Educación, especialidad Inglés. Doctor en Ciencias Pedagógicas. Profesor Titular. Profesor Invitado en Israel Institute of Technology (GTIIT). Shantou. Guangdong, República Popular China. E-mail: manuel.leon@gtiit.edu.cn  ORCID: http://orcid.org/0000-0001-8860-3583

3 Licenciada en Educación, especialidad Inglés. Máster en Desarrollo Social. Doctora en Ciencias de la Educación. Profesora e Investigadora Titular. Centro de Estudios Pedagógicos de la Universidad de Las Tunas, Cuba. E-mail: cariemay@ult.edu.cu, ORCID (https://orcid.org/0000-0002-4414-0772).